This action might not be possible to undo. Are you sure you want to continue?
Molly F. McIntosh • Diana S. Lien • Henry S. Griffis
DPP-2012-U-000503-Final April 2012
The ideas expressed in this paper are those of the authors. The paper does not represent the views of CNA, the Center for Naval Analyses, the Department of the Navy, or the Department of Defense. Distribution unlimited. Copies of this document can be obtained through the CNA Document Control and Distribution Section at 703-824-2123. Copyright 2012 CNA The reproduction of this work for commercial purposes is strictly prohibited. Nongovernmental users may copy and distribute this document in any medium, either commercially or noncommercially, provided that this copyright notice is reproduced in all copies. Nongovernmental users may not use technical measures to obstruct or control the reading or further copying of the copies they make or distribute. Nongovernmental users may not accept compensation of any manner in exchange for copies. All other rights reserved.
Introduction . . . . . . . . . . . . . . . . . . . . . BLS unemployment rates . . . . . . . . . . . . . . “Apples-to-oranges” comparison. . . . . . . . . . . Comparability of unemployment rate denominators . . . . . . . . . . . . . . . . . Comparability of young jobless veterans and nonveterans . . . . . . . . . . . . . . . . A different approach: Comparing unemployment durations . . . . . . . . . . . . . . . . . . . . . . Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 2 3 3 5 6
This page intentionally left blank.
The unemployment rate, an oft-cited labor market statistic, is reported by the Bureau of Labor Statistics (BLS) for veterans and nonveterans.1 Popular press, including the Wall Street Journal and the New York Times, often compares the veteran unemployment rate with the overall national unemployment rate as a way to characterize the civilian economic prospects of veterans.2 In this paper, we discuss the lack of comparability of veteran and nonveteran unemployment rates and consider some of the shortcomings of comparing these metrics. We then suggest an alternative to using unemployment rates to measure the ability of those who leave military service to get a job— namely, how long it takes the respective veteran and nonveteran jobless populations to find employment.
BLS unemployment rates
The BLS uses the Current Population Survey (CPS) to collect information on the employment patterns of U.S. citizens. As part of their monthly data collection, the CPS asks questions on the previous military experience of civilians; anyone who self-reports as having previously been on active duty is considered a veteran. The latest BLS Employment Situation Report from February 2012 reported a veteran unemployment rate of 7.0 percent and a nonveteran unemployment rate of 8.6 percent. 3 The unemployment rate is also reported
1. The authors wish to thank Dr. Russell W. Beland, Deputy Assistant Secretary of the Navy (Military Personnel Policy), for his insights on the veteran unemployment rate, which motivated our initial interest in this topic. See, for example, Tamara Audi, “More States Decide to ‘Buy Veteran,’” Wall Street Journal, Jan. 13, 2012; and Michael A. Fletcher, “Veterans’ Unemployment Outpaces Civilian Rate,” Washington Post, Oct. 16, 2011. The unemployment rate is the share of labor force participants who are currently available for work, unemployed, and have actively sought work in the prior four weeks. The unemployment rate represents a lower bound on the true jobless rate, particularly during recession periods, because it does not include those who are unemployed and discouraged from searching for a job, perhaps owing to the poor job opportunities.
monthly by latest period of military service. Among Gulf War era II veterans (e.g., those serving from September 2001 to present) who are age 18 and older, the February 2012 unemployment rate was 7.6 percent.4 Among these veterans, the unemployment rate has varied over recent years, from a low of 5.6 percent in November 2008 to a high of 15.2 percent in January 2011.5 In addition to gathering monthly labor force status information, the BLS also administers supplemental questions annually on the CPS Veteran Supplement File.6 The results are released in annual Employment Situation of Veterans reports, including annual unemployment rates broken out by age. Unemployment rates by age group are reported on an annual basis, in part because of small-sample-size issues. For 2011, the unemployment rate for young veterans (defined as those age 18 to 24) was 30.2 percent. By comparison, the 2011 unemployment rate for nonveterans of the same age was 16.1 percent.7
A comparison of the BLS veteran and nonveteran unemployment rates is an apples-to-oranges comparison for two reasons. First, the denominators used in the rates are not directly comparable. Second, on the whole, jobless veterans and jobless nonveterans have different characteristics. Here, we discuss these issues in more detail.
BLS, U.S. Department of Labor, The Employment Situation - February 2012, USDL-12-0402, Mar. 9, 2012. BLS, U.S. Department of Labor, Unemployment Rate - Veterans, Gulf War Era II, 18 years and over, 2008 through 2012 monthly data, data series LNU04066408, extracted Apr. 2, 2012. The CPS Veteran Supplement File includes information about servicerelated disabilities, reserve participation, and current issues/concerns, such as transition assistance participation. For more information, see CPS, Jul. 2010, Veterans Supplement File, Technical Documentation CPS10, last accessed Apr. 2, 2012, at http://www.census.gov/apsd/techdoc/ cps/cpsjul10.pdf. BLS, U.S. Department of Labor, Employment Situation of Veterans - 2011, USDL-12-0493, Mar. 20, 2012.
Comparability of unemployment rate denominators
The BLS veteran and nonveteran unemployment rates have comparable numerators; in each case, it is the number of individuals (veterans or not) who are actively seeking work. The denominators for the veteran and nonveteran unemployment rates, however, are not comparable. Specifically, the denominators for the veteran and nonveteran unemployment rates both include individuals who are unemployed. Although the denominator for the nonveteran unemployment rate includes individuals who continue to be employed by civilian employers, the denominator for the veteran unemployment rate does not include the military counterpart to this—namely, individuals who continue to be employed by the military (e.g., individuals on active duty). This difference results in an apples-to-oranges comparison. An “apples-to-apples” comparison would be to define a new “military/ veteran” unemployment rate where the denominator includes current servicemembers as well as all veterans who are currently employed or unemployed. This would amount to adding current servicemembers to the denominator of the BLS veteran unemployment rate.8 While this would allow for an apples-to-apples comparison, it would not be particularly informative with respect to civilian job prospects for veterans.
Comparability of young jobless veterans and nonveterans
The characteristics of young jobless veterans and nonveterans differ in two key ways. First, the pool of young unemployed workers, age 18 to 24, includes those who have recently left their jobs, those who have been without jobs for an extended period of time, and those who
8. By definition, a military/veteran unemployment rate would be lower than the BLS veteran unemployment rate since the only difference between the two is that the military/veteran rate has a larger denominator. Calculation of a military/veteran unemployment rate would be problematic, however, because of differences in data sources (especially differences in the timing of data collection) and the need to account for servicemembers in the reserve component. These complications decrease the practicality of creating an accurate measure of this rate using existing BLS and DOD data.
have never been employed. By definition, a veteran is someone who has been employed in a full-time job. Thus, young nonveterans are more likely than young veterans to have never worked full time. Second, young jobless veterans are potentially less likely than young jobless nonveterans to have a postsecondary education. The typical officer enters the military with a college degree and serves a minimum of four years of obligated active service. Therefore, by the time the typical officer leaves the military, he or she will have aged out of the 18–24 age group. This means that the majority of young veterans will be enlisted veterans, nearly all of which hold only high school degrees. In summary, the pool of young jobless veterans is primarily composed of individuals who have had full-time employment in the past and do not hold a college degree. In contrast, the pool of young jobless nonveterans has a higher share of individuals who have never held a fulltime job and have a postsecondary education. These differences between young jobless veterans and nonveterans makes the BLS young veteran and nonveteran unemployment rates an apples-to-oranges comparison. One alternative would be to compare the unemployment rate for young veterans with the unemployment rate among recent college graduates; individuals in both groups have received postsecondary training and are transitioning from one endeavor (military service or attending college) to another. As noted earlier, however, most young veterans are likely to be high school graduates, and not college graduates. To the extent that the labor market places a different value on military service and a college degree, this limits the usefulness of this comparison. A second alternative would be to calculate a young nonveteran unemployment rate among young nonveterans who recently left their jobs, but this is not a perfect solution either. Not all young jobless veterans recently became unemployed. Some may have left the military early in their military careers because of bootcamp or other training attrition.9
9. For instance, among enlisted servicemembers who joined the military between fiscal years 1999 and 2007, 22.5 percent left within two years. (Lauren R. Malone and Neil Carey, Waivered Recruits: An Evaluation of Their Performance and Attrition Risk, CNA Research Memorandum D0023955.A4, Mar. 2011).
A different approach: Comparing unemployment durations
Given the shortcomings of comparing unemployment rates among veterans and nonveterans, we propose shifting the focus to comparing how long it takes jobless veterans and nonveterans to find employment. This comparison addresses the following policy question: relative to a jobless nonveteran, how hard is it for a jobless veteran to find a civilian job? In a recent report, the Department of Labor published statistics on unemployment durations for data collected in 2010 among jobless veterans and nonveterans.10 For male jobless veterans, the median duration of unemployment was 24.1 weeks compared with 22.7 weeks for male jobless nonveterans. For jobless women, the differences were less pronounced; the median duration of unemployment for female jobless veterans was 21.4 weeks compared with 21.2 weeks for female jobless nonveterans. The report also estimated the share of jobless veterans and nonveterans who were long-term unemployed, defined as unemployment durations of 27 weeks or longer. For men, at the time of data collection in 2010, 47.2 percent of jobless veterans were long-term unemployed compared with 45.2 percent of jobless nonveterans. Again, the differences were smaller for women; at the time of data collection in 2010, 42.5 percent of female jobless veterans were long-term unemployed compared with 42.7 percent of female jobless nonveterans. Finally, the report also breaks out unemployment durations by period of service. For Gulf War era II jobless veterans, at the time of data collection in 2010, the median unemployment duration was 17.8 weeks and 36.1 percent were long-term unemployed. These statistics paint a somewhat rosier picture for Gulf War era II jobless veterans compared with jobless veterans from other periods of service. Still, these statistics do not control for demographic differences across these groups, most notably age.
10. U.S. Department of Labor, The Veteran Labor Force in the Recovery, Nov. 2, 2011.
To get a cleaner comparison, CPS data could be used, if sample sizes are sufficient, to compare unemployment durations across jobless veterans and nonveterans, taking into consideration demographic differences in the two populations, such as gender, age, racial, ethnic, and educational composition.11 In addition, other summary statistics, beyond the median and percent long-term unemployed, could be created. We believe that these more nuanced estimates of unemployment durations for veterans and nonveterans would produce a better understanding of the economic prospects faced by veterans than comparisons of unemployment rates.
While the popular press often cites differences in veteran and nonveteran unemployment rates to characterize the civilian economic prospects of veterans, we posit that these comparisons are tenuous for two reasons. First, the denominators used in the rates are not comparable. Second, jobless veterans and jobless nonveterans have different characteristics. This is especially true for the 18–24 age group, where young jobless nonveterans are more likely to have never held a fulltime job and are more likely to hold college degrees. For these reasons, the BLS veteran and nonveteran unemployment rates are an apples-to-oranges comparison, particularly for the 18–24 age group. There are two ways to make it a straight apples-to-apples comparison. First, the veteran unemployment rate could be redefined as a military/veteran unemployment rate, where the denominator includes current servicemembers as well as all veterans who are currently employed or unemployed. However, this would not be a very informative metric with respect to the civilian job prospects for veterans. Second, for the 18–24 age group, the young nonveteran unemployment rate could be restricted to recent college graduates (who, like young veterans, have received postsecondary training and are transitioning from one endeavor to another) or to those who recently left their
11. Unfortunately, as mentioned earlier with regard to calculating unemployment rates by age group, sample size limitations may restrict the feasibility of this analysis.
jobs. These are likely to fall short of the objective, too. On one hand, the labor market likely values a college education differently from military service; on the other hand, not all young veterans have recently become unemployed. Given the challenges with creating a valid comparison of unemployment rates for veterans and nonveterans, and keeping in mind that the relevant policy issue is whether someone who leaves military service is able to get a job, turning to another metric besides the unemployment rate may be warranted. We posit that unemployment durations are just such a metric. Some aggregate data on unemployment durations for jobless veterans and nonveterans have been reported in a recent Department of Labor report. However, a more nuanced analysis of CPS data is in order—one that considers differences in demographic characteristics between veterans and nonveterans and reports additional statistics beyond the median unemployment duration and the percentage that is long-term unemployed.
This page intentionally left blank.
4825 Mark Center Drive, Alexandria, VA 22311-1850 703-824-2000 www.cna.org
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.